I have a strange exception. I get a "LayoutCycleException" when inserting an item into a ObservableCollection two times with a specific index.
Details: I created a Paging-Class, which handles a big ObservableCollection (up to 2000 items) into a single page (200 items). The user can then go to another page. This event causes the Pager to set the page items new. All works fine except one situation: Inserting new items into the page with specific index (0, 1, 2, ..) fires the exception after the second insert.
Anyone an idee why?
Code:
// Calculate the index for this page
var cp = (CurrentPage == 0) ? 1 : CurrentPage;
int pindex = index - ((cp - 1) * PageItemCount);
if (pindex >= 0 && pindex < PageItemCount)
{
// An item is inserted after the list is already loaded
if (pindex == 0)
{
// Check if the page has mor items than it should
if (_pagedItems.Count >= PageItemCount)
{
// Remove last item from page
_pagedItems.RemoveAt(_pagedItems.Count - 1);
}
_pagedItems.Insert(pindex, item);
}
else if (_pagedItems.Count < PageItemCount)
{
// --> EXCEPTION at third insert
_pagedItems.Insert(pindex, item);
}
}
EDIT (solution)
I found the bug in my code. The Problem was, that the UI had no time to update the collection-changed (UI was not responsive). Very strange bug, took me long time to find it...
A simple Thread.Sleep(10) changed the whole thing =) in this time the ui can change the collection and everything works fine.
The exception is not coming from ObservableCollection, but from an event generated when the layout is updated after InsertItem.
LayoutCycleDetection is typically generated when using the LayoutUpdated event handler. You can unintentionally create an infinite loop. In this case _pagedItems.Insert creates a LayoutUpdate event which may be calling _pagedItems to insert again.
If it's not found there, you may have a custom control calling LayoutUpdated and causing the exception.
Related
I'm trying to figure out a way to build a progress bar with React. I have a forEach loop that iterates through an array of about 7,000 indexes. Each time I validate a row, I want to update a state variable with percentage completion (and render this on the page live). I've tried iterating through these indexes, and updating my state variable (hoping to update the page) in the loop but I'm realizing that will not work. I obviously can't do this with a normal variable as it will reset when the component re-renders. Can anyone give me some insight on this topic?
Thanks.
Here is a code snippet from what I'm looking at:
parsedAssets.forEach(asset => {
newAssetValidated = validateBulkUpload(asset, parsedAssets, assetList, accountLogged, jobSites);
!newAssetValidated.reject_err ? validatedAssetList.accepted.push(newAssetValidated) : validatedAssetList.rejected.push(newAssetValidated);
setStateAssets({ ...stateAssets, validatedAssetList });
});
}
So essentially, as each asset is either accepted or rejected we add it to "stateAssets", and I'm hoping to build the progress bar from the length of the combined arrays that are getting set in stateAssets. However, when the forEach loop is completed, only the last validated asset is getting set due to it not updating until the forEach loop is completed.
Personally I can't imagine such a heavy validation, that you need progress-bar, but anyway.
First solution is to separate validation itself from state update for progress-bar into separate "threads". But since JS is single threaded, you may use some tricks with setTimeout or setInterval functionality. It may be very tricky, and in general not recommended practice with React.
Another way is - to set the work into queue & process 1 item at a time.
As an example I would do it something like this:
function ComponentWithProgress({parsedAssets, setParsedAssets}) {
const [validatedAssetList, setValidatedAssetList] = useState([])
const [progress, setProgress] = useState(0)
const [toDo, setToDo] = useState([])
if(parsedAssets && parsedAssets.length>0) {
setToDo(parsedAssets)
// clear parsedAssets in parent component to: [], false, null ...
// so you put it into toDo only once
setParsedAssets([])
}
if(toDo.length > 0) {
const asset = toDo[0]
const newToDo = toDo.slice(1) // All but 0th element
const newAssetValidated = validateBulkUpload(asset);
setValidatedAssetList([ ...validatedAssetList, newAssetValidated ]);
setToDo(newToDo)
setProgress( newToDo.length / ( validatedAssetList.length + newToDo.length ) * 100 )
}
// ... Render here
// If you need only accepted
const accepted = validatedAssetList.filter(v => !v.reject_err)
}
This example maybe not work for you as is, because you didn't showed us the context, but the main idea is here.
SKAction has waiting for duration abilities, for a period of time on a node. And seems to perform actions on nodes. Like moveTo, etc.
If I don't want that, rather I'd prefer to call functions within GameScene after a period of time, how do I do that with SpriteKit in the GameScene, not on a Sprite or other Node?
Are SKActions the way to do this? The only way to do this?
Yes. This question IS that ridiculously simple. I lack the heuristics and terminology to find an answer. Just keep looping around on how SKAction waits are calls on SKSprites for things like scale, rotation, etc, after time. Which isn't want I want/need.
Update:
Desired outcome, inside GameScene
doSetupStuff() // does some stuff...
waitForAWhile() // somehow wait, perhaps do somethings in here, while waiting
doSomethingElse() // does this after the waitForAWhile has waited
UPDATE 2:
What I think happens, again, inside didMove(to view...)
func wait(){
let timeToPause = SKAction.wait(forDuration: 3)
run(timeToPause)
}
let wontwait = SKAction.wait(forDuration: 3)
run(wontwait)
thisFunction(willnot: WAIT"it starts immediately")
wait()
thisFunction(forcedToWait: "for wait()'s nested action to complete")
UPDATE 3:
Found a way to get the delay without using SKActions. It's a little crude and brutal, but makes more sense to me than SKActions, so far:
DispatchQueue.main.asyncAfter(deadline: .now() + 10.0) {
print("I waited ten seconds before printing this!")
}
An option, as you cited, is to manage this externally. The way I typically manage this sort of thing is to have an externally run update cycle. One that
To drive this updater, you could use either CADisplayLink (which is what I use right now with my OpenGL renderer) or a dispatch source timer (which I have used with my SpriteKit engine). When you use an updated, you want to calculate the delta time. The tick handler could look something like:
func tickHandler() {
let currTime = NSDate().timeIntervalSince1970
let dt = lastTime - currTime // lastTime is a data member of the class
// Call all updaters here, pretend "updater" is a known updater class
updater.update(dt)
}
And updater's update method would look something like:
func update(deltaTime:NSTimeInterval) {
// Do your magic
}
I typically have a main overall updater running independent of what people are calling scenes. Example usage would be something like having an attract mode like in old school arcade games. There they show title screen, sample game play, high scores, rinse and repeat. Scenes would be title, game play, high score. Here you can your main updater manage the time and coordinate the construction/destruction/switching of the scenes. Note this implies having an overall scene manager (which is actually quite handy to have).
For your case, you could use this updater to drive the GameScene updater. It's updater could look something like:
func update(deltaTime:NSTimeInterval) {
switch state {
case .SetupState:
// noop?
println("I'm in setup") // Shown just so you can see there is a setup state
case .WaitState:
waitTime += deltaTime
if waitTime >= kWaitTime {
// Do whats you gots to do
doSomethingElse()
state = .NextState
}
case .NextState:
// blah blah blah blah
}
}
So the flow to do this call path from your driver (CADisplayLink or dispatch source) would be something like:
tickHandler -> master updater -> game scene updater
Some will def find this is perhaps a little heavy handed. I, on the other hand, find this very helpful. While there is obviously some time management and the loss of being able to fire and forget, it can help provide more control for orchestrating pieces, as well as arbitrarily changing state without having to worry about killing already queued actions. There is also nothing that says you still cannot mix SKAction. When I did use SpriteKit, I did all my updating this way along with some dispatched items. I only used SKAction to update hierarchy. Keep in mind that I used my own animation and physics system. So at least for me I had a lot less dependency on SpriteKit (it effectively was just a renderer for me).
Note you have to have your own means to handle pause and coming to foreground where your timer will need to be resynced (you only need to worry about tickHandler). Breakpoints also will cause time jumps.
You can use below function
#define ANIM_TIME 2
SKAction *customACtion = [SKAction customActionWithDuration: ANIM_TIME actionBlock:^(SKNode *node, CGFloat elapsedTime) {
// Do Something Here
}];
Another way to make something happen after a certain period of time is to make use of the 'current time' parm passed to update(). The following code will spawn a boss at intervals ranging from 20 to 30 seconds.
In your property definitions:
var timeOfLastBoss: CFTimeInterval = -1 //Indicate no boss yet
var timePerBoss = CFTimeInterval()
.
.
.
didMoveToView() {
...
timePerBoss = CFTimeInterval(Int.random(20...30))
'''
}
.
.
.
func update(currentTime: CFTimeInterval) {
...
spawnBossForUpdate(currentTime)
...
}
'
'
'
func spawnBossForUpdate(currentTime : CFTimeInterval) {
if ( timeOfLastBoss == -1 ) {timeOfLastBoss = currentTime}
if (currentTime - timeOfLastBoss < timePerBoss) {return}
// Rest of 'spawnBoss code
self.timePerBoss = CFTimeInterval(Int.random(20...30))
self.timeOfLastBoss = currentTime
}
One way, using SKActions, in Swift 3.0, looks like this:
DEFINE: aPatientlyWaitingFunction() at the top level of
GameScene class.
To cause a delay to happen before calling the above function, inside
didMove(to view...)
three ways I've found to do this using Actions:
All three ways seem to accomplish the exact same thing:
let timeToWait: TimeInterval = 3 // is seconds in SKAction thinking time
let waitSomeTime = SKAction.wait(forDuration: timeToWait)
// 1st way __________________________________________
// with a completion handler, the function can be called after Action
run(waitSomeTime) {self.aPatientlyWaitingFunction()}
// 2nd way __________________________________________
// as a completion to be done after action, in the run invocation:
run(waitSomeTime, completion: aPatientlyWaitingFunction)
// 3rd way __________________________________________
// alternatively, as part of a sequence of actions...
// Create a sequence, by making a run action from waitSomeTime and...
let thenDoThis = SKAction.run(aPatientlyWaitingFunction)
// then activate sequence, which does one action, then the next
run(SKAction.sequence([waitSomeTime, thenDoThis]))
// OR... for something different ____________________
////////////////////////////////////////////////////////
DispatchQueue.main.asyncAfter(deadline: .now() + timeToWait) {
self.aPatientlyWaitingFunction()
print("DispatchQueue waited for 3 seconds")
}
I have a webpage(HTML 5) in which there are 4 charts, each of which taking different time to load once the static content in the page comes up. The Loading is shown in the webpage using a 'rendering' circle image for all the 4 charts. I want to find out how much time each of the charts were showing the 'rendering' circle. Please help me in getting a solution using selenium webdriver.
It is the crude way but it must work.
Create a infinite loop with 1 second wait and in each iteration check if the chart is loaded or not. Once all four charts are loaded or if you have timeout value come out of loop.
In this case there is possibility of error of 1 sec. If your chart is loading fast or want to reduce the margin of error reduce the wait from 1 sec to 100msec.
Pseudo code :[May use this to write better code]
boolean[] chartLoaded = {false,false,false,false};
int[] chartLoadTime = {0,0,0,0};
int counter = 0;
while(counter < 100)
{
counter++;
if(isLoaded(chart1))
{
chartLoaded[0] = true;
chartLoadTime[0]++;
}
//Do this for other three charts also
if(chartLoaded[0] && chartLoaded[1] && chartLoaded[2] && chartLoaded[3])
break;
}
Working with Arrays. I create objects on a Class base, push them into an Array, I have 2 buttons: one adds a Child and pushes it into the Array, second one Shifts Array and removes the Child. Also a function on my mouse, if I click an object, I define it's Array number, remove the Child and... well, not sure if successful but "delete Array[i];" where i is target's Array number. I can see Array.length in a text field every time I do something. Second button actually does remove an object from Array, the number decreases. But deleting a specified object from the Array, as well as Array.slice(i,1), doesn't reduce the Array length. So I'm afraid it may cause overflow. It's only Array, not sure, maybe it's fine to have over a million cells in an Array? Like if I make a game with meteor shower, meteors are removed from the screen, but the Array still has their cells. And if they appear like 30-50 per sec, it's obvious I may get memory problems in 20 minutes of running it. Well it's 60k so maybe I shouldn't worry as only graphics take much memory?
Still, I could use an advice on how to shift an object in a middle of an Array. Chosen one. How do I delete it as if it never was created, same as Shift does? (it does, right?)
Array.splice() will do most of the tricks. As you remove the meteor off screen, you can splice it out of the array, do like this:
var i:int=meteorArray.indexOf(meteorToRemove);
if (i>=0) meteorArray.splice(i,1);
You should not create a new object as a good practice if you are deleting other. Try recycling, or better said 'object pools'
You should figure out the number max of elements you may use at once. create a property 'active' on each and set it to true or false instead of creating/deleting. Then you can run the update on each object and update it only if necessary
For example:
var meteors:Array = [];
// create 500 meteors
for ( var i:int = 0; i < 500; i++ ) {
var meteor:Meteor = new Meteor();
meteor.active = false;
meteors.push(meteor);
}
// enable one meteor
meteors[0].active = true;
// in your update method:
for each (var meteor in meteors )
if ( meteor.active )
meteor.update();
Hope that helps.
Also you can add a helper method to get a meteor available:
function getMeteorAvailable():Meteor
{
for each (var meteor in meteors )
if ( !meteor.active )
return meteor;
}
I'm trying to create a flash application that will keep track of user generated values. The app should basically allow the user to input the name of the item and it's cost. The total costs should then be added up to show a total value to the user. I can probably figure out how to add the values together, but I'm not really sure how to allow the user to create a list and then allow the user to save it. Can anyone point me towards a tutorial or point me in the right direction?
I am using variables to add user inputed numbers to come up with a total. The first problem is that actionscript 3.0 does not allow variables for texts. I just converted it to 2.0 to fix this. The second problem, is when I test the app and put in my values and click submit, I get NaN in the total values field. Is there a reason why it wouldn't add the values?
Here is the code I used for the submit button:
on (release) {
total = Number(rent) + Number(food) + Number(travel) + Number(entertainment) + Number(bills);
}
Am I missing anything?
Can I give the input text instance names and then give them variables? How are some ways to go about this?
Thanks for the help!
Have an object array, say for example
var stack:Array = new Array();
Then push the item name and it's cost to that array when user inputs, like
stack.push({item:AAA, cost:xx});
So that you can generate the list whenever you want with that array.
You have to see how this works in code. A list in actionscript could be stored inside an array, vector, dictionary or even an Object.
Var myList:Array = [];
myList.push({name: "item 1", cost: 5 });
myList.push({name: "item 2", cost: 7.5 });
If you want to grab the 'product' of "item 1" from the list, you have to create a function for that, lets call it getProductByName
function getProductByName(name:String):Object
{
for each(var product:Object in myList)
{
if (product.name === name) return product;
}
return null; // no match found
}
You can call that function like this:
var product = getProductByName("item 1");
trace(product.cost); // 5
And you can alter the product, so lets make it more expensive
product.cost += 1;
trace(product.cost); // 6
Have fun! If you are using classes, you would create one for the product, with public name and cost, and in that case you'de better use a vector, to ensure working with the right type.
This is what fixed the issue for me in action script 3.0:
myButton.addEventListener(MouseEvent.CLICK, addThem);
function addThem(e:MouseEvent)
{
totalField.text = String ( Number(field1.text) + Number(field2.text) + ....);
}
I also had to name the instances appropriately.